Triple-negative breast cancer: can spatial and single-cell profiling predict response to immunotherapy and radiation?

By Ana Espino | Published on october 21, 2025 | 3 min read


Triple-negative breast cancer (TNBC), lacking hormonal and HER2 receptors, represents an aggressive and heterogeneous subtype. Immune checkpoint inhibitors (ICIs), such as pembrolizumab, have recently shown therapeutic potential; however, only about 20% of metastatic patients respond to monotherapy. This highlights the urgent need for predictive biomarkers and more effective combination strategies.

An aggressive subtype in search of new approaches

This study aimed to decipher the spatial and cellular immune dynamics underlying response to pembrolizumab and neoadjuvant radiation therapy (RT). Using single-cell RNA sequencing (scRNA-seq) and spatial proteomic mapping (CODEX) on longitudinal biopsies, the authors identified three distinct response trajectories in patients with TNBC.  

How does immunity respond to TNBC? Three trajectories revealed by spatial mapping 

The clinical trial (NCT03366844) enrolled 50 newly diagnosed TNBC patients. The protocol included two cycles of pembrolizumab, the second combined with focal radiotherapy (24 Gy in 3 fractions), followed by standard treatment (chemotherapy and surgery). Biopsies were performed before treatment, after pembrolizumab alone, and after the pembrolizumab + RT combination. Among the 34 patients with complete follow-up, 67% responded to treatment (pCR or RCB score 0–1). Immunological data revealed three trajectories:

R1: Early responders (pre-immunized phenotype)

This group is characterized by:

• High lymphocytic infiltration from the outset (high TILs)
• Presence of tertiary lymphoid structures (TLS)
• High MHC I/II expression by tumor cells
• Rapid expansion of effector CD8+ cells after pembrolizumab

These patients exhibit an “inflamed” phenotype with pre-existing immunity and respond to pembrolizumab alone, with no clear additional benefit from radiotherapy (RT).

R2: Late responders (radiation-induced immunity)

Initially similar to non-responders, these tumors develop a strong immune response only after the pembrolizumab + RT combination:

• Recruitment of antigen-presenting macrophages (sc-m11)
• Expansion of effector CD8+ cells (sc-t6)
• Activation of migratory dendritic cells (sc-m7)
• Increased spatial proximity between T cells, macrophages, and tumor cells

This profile corresponds to an “excluded” phenotype — refractory to ICI alone but activated by RT, which enhances antitumor immunity.

NR: Non-Responders (Immune-Ignored Phenotype)

This group is characterized by:

• A persistent lack of immune infiltration
• Low expression of MHC-related genes
• Little to no change in the tumor microenvironment after treatment

Their immunologically silent (“ignored”) phenotype remains refractory to all tested treatment combinations. 

Read next: Triple blow, triple clinical challenge

Can immunotherapy be personalized according to the tumor’s spatial profile?

This study demonstrates that TNBC is not immunologically uniform. Two response profiles can be observed:

• A pre-immunized group (R1), responsive to ICI alone
• A “cold” but activatable group (R2), which clearly benefits from the addition of RT

In contrast, the non-responder group (NR) highlights the existence of a biological barrier that remains unovercome. These findings emphasize the importance of spatial and single-cell profiling prior to treatment to tailor therapeutic combinations.

Finally, the murine model confirmed the synergistic potential of RT and pembrolizumab in activating differentiated CD8+ T cells and MHC-II+ macrophages, supporting the clinical translation of these observations.

Perspectives: integrate these data to stratify patients — considering chemotherapy de-escalation in R1 or targeted intensification in R2. The future lies in personalized immuno-spatial medicine.

Read next: TNBC: could everything be happening in the mitochondria?